Merge pull request #409 from Bartoshr/bugfix/short-channel-urls

Support short custom YouTube channel URLs
This commit is contained in:
Stypox 2020-10-25 11:43:50 +01:00 committed by GitHub
commit 8026304a0a
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 43 additions and 7 deletions

View file

@ -1,5 +1,6 @@
package org.schabi.newpipe.extractor.services.youtube.linkHandler;
import java.util.regex.Pattern;
import org.schabi.newpipe.extractor.exceptions.ParsingException;
import org.schabi.newpipe.extractor.linkhandler.ListLinkHandlerFactory;
import org.schabi.newpipe.extractor.services.youtube.YoutubeParsingHelper;
@ -32,6 +33,9 @@ public class YoutubeChannelLinkHandlerFactory extends ListLinkHandlerFactory {
private static final YoutubeChannelLinkHandlerFactory instance = new YoutubeChannelLinkHandlerFactory();
private static final Pattern excludedSegments =
Pattern.compile("playlist|watch|attribution_link|watch_popup|embed|feed|select_site");
public static YoutubeChannelLinkHandlerFactory getInstance() {
return instance;
}
@ -48,11 +52,22 @@ public class YoutubeChannelLinkHandlerFactory extends ListLinkHandlerFactory {
public String getUrl(String id, List<String> contentFilters, String searchFilter) {
return "https://www.youtube.com/" + id;
}
/**
* Returns true if path conform to
* custom short channel URLs like youtube.com/yourcustomname
*
* @param splitPath path segments array
* @return true - if value conform to short channel URL, false - not
*/
private boolean isCustomShortChannelUrl(final String[] splitPath) {
return splitPath.length == 1 && !excludedSegments.matcher(splitPath[0]).matches();
}
@Override
public String getId(String url) throws ParsingException {
try {
URL urlObj = Utils.stringToURL(url);
final URL urlObj = Utils.stringToURL(url);
String path = urlObj.getPath();
if (!Utils.isHTTP(urlObj) || !(YoutubeParsingHelper.isYoutubeURL(urlObj) ||
@ -60,15 +75,21 @@ public class YoutubeChannelLinkHandlerFactory extends ListLinkHandlerFactory {
throw new ParsingException("the URL given is not a Youtube-URL");
}
if (!path.startsWith("/user/") && !path.startsWith("/channel/") && !path.startsWith("/c/")) {
// remove leading "/"
path = path.substring(1);
String[] splitPath = path.split("/");
// Handle custom short channel URLs like youtube.com/yourcustomname
if (isCustomShortChannelUrl(splitPath)) {
path = "c/" + path;
splitPath = path.split("/");
}
if (!path.startsWith("user/") && !path.startsWith("channel/") && !path.startsWith("c/")) {
throw new ParsingException("the URL given is neither a channel nor an user");
}
// remove leading "/"
path = path.substring(1);
String[] splitPath = path.split("/");
String id = splitPath[1];
final String id = splitPath[1];
if (id == null || !id.matches("[A-Za-z0-9_-]+")) {
throw new ParsingException("The given id is not a Youtube-Video-ID");

View file

@ -8,6 +8,7 @@ import org.schabi.newpipe.extractor.exceptions.ParsingException;
import org.schabi.newpipe.extractor.services.youtube.linkHandler.YoutubeChannelLinkHandlerFactory;
import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertFalse;
import static org.junit.Assert.assertTrue;
/**
@ -30,6 +31,8 @@ public class YoutubeChannelLinkHandlerFactoryTest {
assertTrue(linkHandler.acceptUrl("https://www.youtube.com/c/creatoracademy"));
assertTrue(linkHandler.acceptUrl("https://youtube.com/DIMENSI0N"));
assertTrue(linkHandler.acceptUrl("https://www.youtube.com/channel/UClq42foiSgl7sSpLupnugGA"));
assertTrue(linkHandler.acceptUrl("https://www.youtube.com/channel/UClq42foiSgl7sSpLupnugGA/videos?disable_polymer=1"));
@ -44,6 +47,18 @@ public class YoutubeChannelLinkHandlerFactoryTest {
assertTrue(linkHandler.acceptUrl("https://invidio.us/channel/UClq42foiSgl7sSpLupnugGA"));
assertTrue(linkHandler.acceptUrl("https://invidio.us/channel/UClq42foiSgl7sSpLupnugGA/videos?disable_polymer=1"));
assertTrue(linkHandler.acceptUrl("https://www.youtube.com/watchismo"));
// do not accept URLs which are not channels
assertFalse(linkHandler.acceptUrl("https://www.youtube.com/watch?v=jZViOEv90dI&t=100"));
assertFalse(linkHandler.acceptUrl("http://www.youtube.com/watch_popup?v=uEJuoEs1UxY"));
assertFalse(linkHandler.acceptUrl("http://www.youtube.com/attribution_link?a=JdfC0C9V6ZI&u=%2Fwatch%3Fv%3DEhxJLojIE_o%26feature%3Dshare"));
assertFalse(linkHandler.acceptUrl("https://www.youtube.com/playlist?list=PLW5y1tjAOzI3orQNF1yGGVL5x-pR2K1d"));
assertFalse(linkHandler.acceptUrl("https://www.youtube.com/embed/jZViOEv90dI"));
assertFalse(linkHandler.acceptUrl("https://www.youtube.com/feed/subscriptions?list=PLz8YL4HVC87WJQDzVoY943URKQCsHS9XV"));
assertFalse(linkHandler.acceptUrl("https://www.youtube.com/?app=desktop&persist_app=1"));
assertFalse(linkHandler.acceptUrl("https://m.youtube.com/select_site"));
}
@Test